HYPERSPECTRAL PRISMA DATA PROCESSING FOR WATER QUALITY RESEARCH AND APPLICATIONS

Climate change is having a significant negative impact on freshwater systems, which provide multiple ecosystem services. In this context, the present study aims to show an overview of the main objectives achieved by exploiting the hyperspectral reflectance data provided by the PRISMA sensor to map aquatic ecosystems. Water quality products were generated using three different approaches: the biooptical model BOMBER, adaptive semi-empirical algorithms, and machine learning models. These methods were tested in very different waterbodies worldwide: five lakes, two lagoons and one river. To assess the accuracy of the water quality products, comparisons were performed with reference measurements. The results showed an average R2 = 0.70 and encourage using PRISMA data for aquatic applications in synergy with existing multispectral and future hyperspectral data.